We have started using the Irrometer 200SA sensor in one of our products. We are using an AVR series micro-controller and we have followed the hardware instructions provided by the Irrometer team. We have used Multiplexer to use more than 1 sensor in a device.
Now as per our testing, we are getting a good response i.e. we can clearly see the changes when the irrigation is done ( screenshot attached ).
Now the problem i am facing is % changes in the readings while in the same environment ( we are sensing every 10 minutes ). In the same environment, the resistance which we are calculating is changing up to 9% also ( below is the reading and corresponding changes ). I have few questions
If you convert the resistance to CB ( centibar - this is what any agronomist will use ), the change in CB is negligible ( using their calibration lookup table) as per our resistance values.
Is the sensor readings suppose to change this much within 10 minutes?
Irrometer guys have not published any resolution or accuracy. This means there is no way, I can find out if the resistance values can vary this much % between corresponding readings. There is no data on repeatability published by them.
Does anyone else has used these sensors and came across these issues of % changes in the corresponding readings even in almost the same soil conditions?
We got the schematic and code snippets verified by the Irrometer team and they say, they are fine. Though they mentioned that the variations in sensor readings is a bit outside the expected one. This is where my confusion lie. Irrometer has not published any accuracy specs of the sensor.
We measure the resistance which is then converted into kPa ( by the equation they have provided in the sensor page ) or into CB ( using the resistance vs CB lookup table ).
If we draw a scatter plot, the variation seems to be ok ( the maximum variation between successive readings sometime is +/-7-8% ).
Hi,
The problem is you can't just connect a fixed resistor across the sensors and check for stability/drift.
A suggestion would be to measure a sample of soil that is sealed from the atmosphere to prevent evaporation and measure its drift.
If your change in CB readings are negligible, it sounds like some resistance difference is tolerable.
Is this what you are worried about?
The noise?
Yes, the noise is what I am worried about. Though I understand that any resistive sensor would have some accuracy and no sensor will give the exact readings. I would have been happier if the sensor manufacturer would have published the accuracy data.
Below is the image which shows the Kohm, kPa vs time graph.
I am ok with the performance of the sensor i.e. we are able to measure the time of irrigation and able to figure out when the soil will need irrigation and that fulfills the sensor usage.
Hi,
That noise I think you will find will always be there and as you are using CB and the noise does not appear to have an effect on change, you may have to be happy with what you have got.
This statement;
"Devices reading WATERMARK sensors must isolate the sensor reading circuit from any earth ground. "
Seems to me to say it is susceptible to noise and any earth currents.
